Image fusion is to use a certain algorithm to comprehensively process two or more different source images, and finally form a new image. According to the different stages of fusion processing, image fusion processing can usually be performed at three different levels: pixel-level image fusion, feature-level image fusion, and decision-level image fusion. ) image fusion.

 

Foreign countries have carried out research on low-light and infrared image fusion algorithms for a long time, and have proposed a series of basic algorithms, among which the pixel-level methods mainly include weighted average method, pyramid fusion method, and principal component analysis (PCA). ) method, multi-resolution tower algorithm, wavelet transform method, high-pass filter method, Kalman filter algorithm, regression model method, etc.